Redact PDF
Permanently remove content by rasterising pages — not a black box over live text.
Drawing a black rectangle over text does not remove it. The text stays in the file and can be copied straight out from underneath — a mistake that has exposed information in court filings and government releases more than once. This tool removes content properly, by rebuilding every page as an image.
Your file never leaves your device. It is read and edited by JavaScript running in this tab. Nothing is uploaded or stored.
Enter your details
Result
Fill in the fields above and your result will appear here.
What the Redact PDF does
Each page is rendered to a canvas, your black areas are painted on, and the result is embedded as an image in a new PDF. Because the page is rebuilt from pixels, the original text objects no longer exist anywhere in the output — there is nothing left to recover.
The cost is real and worth understanding: the whole document becomes images. Text is no longer selectable, searchable or readable by screen readers, and the file may be larger than the original.
How to use this tool
- Check where the sensitive content sits, measuring roughly as a percentage across and down the page.
- Enter each area as
x,y,width,heightin percentages, separated by semicolons. - Set which pages to apply the boxes to. Every page is flattened regardless, which is what makes the redaction safe.
- Press Redact and flatten, then verify the result with the text extractor — it should find nothing.
Why flattening is what makes it safe
UNSAFE (what most tools do)
Draw a black rectangle over the text.
→ the text object is still in the file
→ copy-paste, or any extractor, recovers it instantly
SAFE (what this does)
1. Render each page to pixels.
2. Paint the black areas onto those pixels.
3. Rebuild the PDF from the images.
→ no text objects survive anywhere in the document
→ there is nothing underneath to recover
Coordinates are percentages of the page: 0,0 is top-left.
Worked example
Example: hiding an account number
| Area entered | 35,28,40,4 |
| Meaning | 35% across, 28% down, 40% wide, 4% tall |
| Applied to | pages 1-2 |
| Verification | text extractor finds no text at all |
Always verify. Running the output through the text extractor and finding nothing is the proof that redaction worked.
What your result means
The whole document is flattened, not just the pages with boxes. That is deliberate — partial flattening would leave recoverable text elsewhere in the file.
Verify every time. Run the result through the text extractor; it should report no text found.
The file may be larger, especially if the original was a text document. That is the price of removing the text.
Important considerations
- Positioning is by percentage rather than by clicking, so check the output and adjust if a box is misplaced. Err on the side of larger boxes.
- Metadata is not content. Run the metadata remover as well before sharing anything sensitive.
- Redaction is irreversible in the output, which is the point. Keep your original somewhere safe.
- Cropping is not redaction — cropped content still exists in the file and can be restored.
- For legal disclosure obligations, follow whatever process your jurisdiction or client requires; this is a technical tool, not legal advice.
- Accessibility is lost, since screen readers cannot read a flattened page.
Limitations of this tool
- Password-protected PDFs cannot be opened without the password.
- Processing happens in your device's memory, so very large documents can be slow.
- Damaged or partially downloaded files cannot be parsed.
- Areas are specified numerically rather than drawn with the mouse.
- The entire document is rasterised, so text becomes unsearchable and inaccessible to screen readers throughout.
- Embedded attachments and JavaScript are not carried across, which is usually desirable here but worth knowing.
Frequently asked questions
Is drawing a black box over text safe?
No, and this is the most common redaction failure. The text remains in the file underneath and can be copied out or extracted in seconds. Only removing the underlying content is safe.
How do I know the redaction worked?
Run the output through the text extractor. If it reports no text found, the text objects are gone.
Why is the whole document flattened?
Because partial flattening would leave recoverable text on the pages that were not rasterised. Flattening everything is what makes the guarantee hold.
Can redacted content ever be recovered?
Not from the output file — the text objects no longer exist. It can obviously still be recovered from your original, so store or delete that deliberately.